The Final Two Silos

It is interesting to see the Enterprise space, with all the efforts of integration and collaboration, continue to struggle with fundamental technology differences and the lack of willingness to improve. There seem to be two final silos of not just data, but process and infrastructure. The silos are between historically perceived enterprise holdings and new technology enterprise holdings.

1. The database. Most large enterprises seem to have heavy investments in Oracle and MSSQL. Full disclosure on my preference, but there are valid reasons for that.
2. The programming platform. Most large enterprises have a heavy investment between Java and .Net with a sprinkling of other things here and there for one off. IBM comes into play here as well with its horrid product line.
3. Enterprise Content Management. Every large enterprise is working through this challenge. The two major camps that I see left are EMC and Microsoft. The file server is dead and if you are still using one, well, quit that … better yet, call me and I’ll tell you how to quit that and save 80% of your TCO but I digress.
4. Collaboration space. The two major ones I see out there are still the legacy backed DBs like SAP/Oracle/IBM and the newer SharePoint space.

These two paths boil down to this: Oracle and Java vs Microsoft and .Net. Integration between the two is historically painful, despite some honest to goodness attempts by Oracle and MS to offer cross platform dev tools. The philosophical differences are profound here and are highlighted using the above examples and detailed here:

1. Monolithic black box systems and licensing models that border on the insane – read Oracle/SAP/EMC things that are just hard to use for no apparent benefit and of course the ludicrous recurring maintenance fees.
2. Lighter weight open systems with slightly less painful licensing models that are much easier to afford – read Microsoft things that scale out (in Windows world granted) with no recurring maintenance fees but a complex seizure educing original licensing structure.

Justification for my thought: Oracle’s recent purchase of Sun and its subsequent spike of maintenance fees for legacy systems as well as their own internal model of licensing portends a huge gap between the reality of the modern enterprise and what IT or BDMs think the reality is. Put simply – Oracle/IBM are trying desperately to remain relevant by hawking new age front ends to old school cumbersome products that are impossible to extend or use. The opposing force is Microsoft with its admittedly complex issues of installation and maintenance (why I have to still reboot once a month for freaking patches is beyond me) but ease of end user adoption, functionality for both end users and administrators that doesn’t look like a two year old (or an engineer) designed it and amazingly lower TCO.

This post was a bit of a ramble and I’ll try to articulate better soon, but suffice to say, I have a major bias but one that appears valid, at least from my vantage point.